Google team publishes EmbeddingGemma 2: unified multimodal embeddings for text, image, audio, video
_reachsumit · x · 2026-10-29
A large Google team (Henrique Schechter Vera, Sahil Dua, et al.) released the EmbeddingGemma 2 paper on arXiv. It presents an open embedding model that maps text, images, audio, and video into one shared representation space, with modular encoders designed for smaller deployments.
Related event: Google Open-Sources EmbeddingGemma 2, a 744M Omni-Modal Embedding Model(13 posts)→
More from Multimodal
- Underdog Woof: 31.7MB open-weight TTS model claims 1.42% WER, runs fully in-browser — _sathvikr · 2026-11-06
- CVC presents 'Albedo Estimation via Latent Bridge Matching' at CIC 2026 with open code and weights — anand_bhattad · 2026-11-06
- Vimoo 1.0-Lite open-sourced: single 9.5B model for both text-to-video and image-to-video — LatentSpacer · 2026-11-06
- Krea launches browser-based Krea Video with cinematic controls and 100+ presets — angrypenguinPNG · 2026-11-06
- P-JiT 21K: 2B pixel-space text-to-image model trained for $21K gets native ComfyUI port — Apprehensive_Sky892 · 2026-11-06
- User shows 100 consistent voxel 3D assets reportedly generated by GPT-6.1 Sol — Angaisb_ · 2026-11-06